Fast Model Editing at Scale

by   Eric Mitchell, et al.

While large pre-trained models have enabled impressive results on a variety of downstream tasks, the largest existing models still make errors, and even accurate predictions may become outdated over time. Because detecting all such failures at training time is impossible, enabling both developers and end users of such models to correct inaccurate outputs while leaving the model otherwise intact is desirable. However, the distributed, black-box nature of the representations learned by large neural networks makes producing such targeted edits difficult. If presented with only a single problematic input and new desired output, fine-tuning approaches tend to overfit; other editing algorithms are either computationally infeasible or simply ineffective when applied to very large models. To enable easy post-hoc editing at scale, we propose Model Editor Networks with Gradient Decomposition (MEND), a collection of small auxiliary editing networks that use a single desired input-output pair to make fast, local edits to a pre-trained model. MEND learns to transform the gradient obtained by standard fine-tuning, using a low-rank decomposition of the gradient to make the parameterization of this transformation tractable. MEND can be trained on a single GPU in less than a day even for 10 billion+ parameter models; once trained MEND enables rapid application of new edits to the pre-trained model. Our experiments with T5, GPT, BERT, and BART models show that MEND is the only approach to model editing that produces effective edits for models with tens of millions to over 10 billion parameters. Implementation available at


page 1

page 2

page 3

page 4


Memory-Based Model Editing at Scale

Even the largest neural networks make errors, and once-correct predictio...

KNN-BERT: Fine-Tuning Pre-Trained Models with KNN Classifier

Pre-trained models are widely used in fine-tuning downstream tasks with ...

Adaptive Budget Allocation for Parameter-Efficient Fine-Tuning

Fine-tuning large pre-trained language models on downstream tasks has be...

BERT Fine-Tuning for Sentiment Analysis on Indonesian Mobile Apps Reviews

User reviews have an essential role in the success of the developed mobi...

Editing Models with Task Arithmetic

Changing how pre-trained models behave – e.g., improving their performan...

Robustness of edited neural networks

Successful deployment in uncertain, real-world environments requires tha...

My3DGen: Building Lightweight Personalized 3D Generative Model

Our paper presents My3DGen, a practical system for creating a personaliz...

Please sign up or login with your details

Forgot password? Click here to reset